Trade data headlines show the trade balance modestly worsened with both imports and exports increasing.

Analyst Opinion of Trade Data

From the BEA:

Exports and imports in September reflect both the ongoing impact of the COVID-19 pandemic and the continued recovery from the sharp declines earlier this year. The full economic effects of the pandemic cannot be quantified in the trade statistics because the impacts are generally embedded in source data and cannot be separately identified. The Census Bureau and the Bureau of Economic Analysis continue to monitor data quality and have determined estimates in this release meet publication standards.

The data in this series wobbles and the 3-month rolling averages are the best way to look at this series. The 3-month average rate of growth improved for imports and exports - but remains in contraction.

Econintersect uses the import trade data as a factor in determining the acceleration or deceleration of the economy - but does not believe the negative trade balance per se is an economic issue.

Note that the headline numbers are not inflation-adjusted. Taking a step back and looking at inflation-adjusted data, it is in contraction for both exports and imports.

• Headlines said Imports of goods were up month-over-month - import goods growth has positive implications historically to the economy. Econintersect analysis shows unadjusted goods (not including services) growth accelerated 0.2 % month-over-month (unadjusted data) - up 0.0 % year-over-year (up 1.1 % year-over-year inflation-adjusted). The rate of growth 3-month trend improved.
• Headlines said Exports of goods were up month-over-month, and Econintersect analysis shows unadjusted goods export growth accelerated (not including services) 2.5 % month-over-month - down 7.0 % year-over-year (down 5.4 % year-over-year inflation-adjusted). The 3-month rate of growth trend improved.

• The market expected (from Econoday) a trade balance of $-67.1 B to $-63.5 B (consensus $64.8 B billion deficit) and the seasonally adjusted headline deficit from US Census came in at $63.1 billion.
• It should be noted that oil imports were up 5 million barrels from last month, and down 25 million barrels from one year ago.
• The data in this series is noisy, and it is better to use the rolling averages to make sense of the data trends.

The headline data is seasonally but not inflation-adjusted. Econintersect analysis is based on the unadjusted data, removes services (as little historical information exists to correlate the data to economic activity), and inflation adjusts. Further, there is some question whether this service portion of export/import data is valid in real-time because of data gathering concerns. Backing out services from import and exports shows graphically as follows:

Growing exports is a sign of an expanding global economy (or at least a sign of growing competitiveness).

Seasonally Adjusted Total Imports (blue line), Exports (red line), and Trade Balance (green line)

Econintersect is most concerned with imports as there is a clear recession link to import contraction. Adjusting for cost inflation allows apples-to-apples comparisons in equal value dollars between periods. The graph below uses seasonally adjusted data.

Seasonally and Inflation-Adjusted Year-over-Year Change Imports (blue line) and Exports (red line)

Note: In general this is a rearview look at the economy - however, imports do have a forward vision of up to three months ahead of expected economic activity.

Caveats on Using this Trade Data Index

The data is not inflation-adjusted. Econintersect applies the BLS export-import price indices to the data to adjust for inflation - total exports, total imports, and imports less oil. Adjusting for cost inflation allows apples-to-apples comparisons in equal value dollars between periods.

Although Econintersect generally disagrees with the seasonal adjustment methodology of the U.S. Census, in general, this methodology works for this trade data series as the data is not as noisy as other series. Another positive aspect of this series is that backward revision has usually been very minor.

Econintersect determines the month-over-month change by subtracting the current month's year-over-year change from the previous month's year-over-year change. This is the best of the bad options available to determine month-over-month trends - as the preferred methodology would be to use multi-year data (but the New Normal effects and the Great Recession distort historical data).

Oil prices, and also quantities of imported oil, wobble excessively year-over-year and month-over-month. In 2010, the percent of oil imports varied between 10.4% and 14.6% of the total. In 2008 the variance was between 11.5% to over 20%. No amount of adjusting - short of removing oil imports from the analysis - allows a clear picture of imports.

An improving trade balance historically is a sign of a slowing economy.